Repair or replace? A decision framework that is not just price

The usual advice is to compare the repair quote against the replacement price and pick the cheaper one. That is a reasonable starting point and a bad stopping point, because it ignores how long the repaired device will remain useful.

Here is a framework that accounts for that.

Step one: the 50 percent rule

If the repair costs more than half of what an equivalent replacement costs, repairing is usually the worse deal. Not always, but usually, and it is a fast filter.

Two numbers you need to get right. The repair cost should include shipping and any diagnostic fee that is not credited against the work. The replacement cost should be the price of a comparable device, not the flagship you would like, and should account for what you could recover selling or trading in the broken unit.

For most phones, an out-of-warranty screen replacement sits well under half the cost of a new equivalent, which points to repair. A logic board replacement often does not.

Step two: remaining software support

This is the step people skip, and it is frequently decisive.

A device that no longer receives security updates is a liability regardless of whether the hardware works. Check the manufacturer’s published support window for your model before spending anything.

As a rough guide: major phone manufacturers now commit to several years of OS and security updates from release, and the specific commitment varies significantly by brand and by model tier. Look up your exact model rather than assuming.

Step three: the parts availability test

Some repairs are cheap now and impossible in two years. Before repairing an older device, check whether parts are still being produced. If replacement parts are already scarce or only available from salvaged units, a repair buys you less time than the numbers suggest.

This cuts the other way too. For devices with strong parts availability and an active independent repair market, repair is often the clearly better economic choice long past the point where the 50 percent rule alone would suggest replacing.

Step four: what actually failed

Failures are not equal. Ranked from most to least worth repairing:

Usually worth repairing. Screens, batteries, charging ports, keyboards, fans, single failed RAM modules or drives. These are discrete, well-understood parts with predictable costs, and fixing one does not signal that others are about to fail.

Sometimes worth repairing. Cameras, speakers, hinges, and anything requiring partial disassembly of a bonded assembly. Cost varies wildly by model.

Rarely worth repairing. Logic board and motherboard failures, water damage affecting multiple components, and structural damage to the chassis. These are expensive and often indicate the device has taken damage beyond the visible fault.

The DIY question

Self-repair changes the maths considerably. A battery you pay $120 to have replaced might cost $35 in parts.

It is worth attempting if the device has a published repair guide, if the repair does not require a heat gun and adhesive work, and if failure would not destroy a device that currently mostly works. Batteries in devices with pull-tab adhesive, laptop RAM and storage, and desktop components are all reasonable first repairs.

It is worth avoiding if the display is bonded to the glass, if the repair requires board-level soldering, or if the device is your only phone and you need it working tomorrow.

If you do go this route, a proper toolkit with the right drivers, spudgers and suction handles matters more than most people expect. Cheap kits round off screw heads on the first attempt.

Putting it together

Repair when: the fault is discrete, parts are available, more than two years of software support remain, and the cost is under half of replacement.

Replace when: the board has failed, support is ending, parts are scarce, or the device has multiple faults at once.

The last one is underrated. A device with a failing battery and an intermittent port and a cracked back is telling you something about its remaining life, and fixing one of the three does not change the message.